Summary

The discussions about the spurious directions in triclinic anisotropy by Vavryčuk are not completely accurate. Theoretically, some true acoustic axes may be eliminated by mistake when applying rules proposed. In this paper, correct equations for determining spurious directions are given. It is also pointed out that despite the theoretical flaws in the paper, the subsequent numerical tests and discussions are still valid.

The existence of acoustic axes (singularities, degeneracies) is a common feature in anisotropic media. Along acoustic axes, the phase velocities of two or three plane waves coincide, and the behaviour of the degenerate waves becomes complicated in their vicinity. To calculate the acoustic axes in triclinic anisotropy, Vavryčuk (2005) solves the following Khatkevich equations (see Khatkevich 1963, eq. 11; Vavryčuk 2005, eqs 6a–c):
(1a)
(1b)
(1c)
with Γik = aijklnjnl the Christoffel tensor, aijkl the density-normalized stiffness tensor and nj the wave normal. However, only two of eqs (1a)–(c) are independent and solving them yields at most 36 isolated directions, among which only 16 directions are true acoustic axes, while the other 20 are spurious and should be eliminated. Considering that the Khatkevich equations are derived via multiplying the Darinskii equations by terms Γ12Γ13, Γ12Γ23 or Γ13Γ23, Vavryčuk (2005) claims that the solutions of equations (see his eqs 7a–c)
(2a)
or
(2b)
or
(2c)
are the spurious directions and should be eliminated from the complete set of solutions of the Khatkevich equations. Unfortunately, it is not accurate to call all the solutions of eqs (2a)–(c) spurious and simply eliminate them without further considerations. As shown below, not all the solutions of eqs (2a)–(c) must be necessarily spurious.
If two eigenvalues of the Christoffel tensor coincide (i.e. a P-S1 singularity or a S1-S2 singularity), the solution of eqs (2a)–(c) can be a true acoustic axis if one of the following conditions is satisfied:
3
or
(4a)
or
(4b)
or
(4c)

If all of Γ12, Γ13 and Γ23 vanish and Γ11 = Γ22 ≠ Γ33 (see eq. 3), then G = Γ11 = Γ22 is the degenerate eigenvalue. As an example, we can mention anisotropy with elastic constants a34 = a35 = a45 = 0 and a33a44 = a55. The acoustic axis is along the x3 axis and it is called the longitudinal acoustic axis (Fedorov 1968) since it coincides also with the longitudinal direction.

If only two of Γ12, Γ13 and Γ23 vanish, for example, Γ13 = Γ23 = 0 and Γ12 ≠ 0 (see eq. 4a), the three eigenvalues of the Christoffel tensor Γik are given by
5
If the last identity in eq. (4a) is satisfied, one obtains that
6
Substituting eq. (6) into eq. (5), one obtains that Γ33 is the degenerate eigenvalue and Γ11 + Γ22−Γ33 is the non-degenerate one. As an example, we can mention anisotropy which possesses a longitudinal acoustic axis not coinciding with any of the coordinate axes but lying in a coordinate plane.
If three eigenvalues of the Christoffel tensor coincide in some direction (the so-called triple degeneracy), all three plane waves have identical phase velocities and the Christoffel tensor is isotropic:
7
Obviously, eqs (2a)–(c) are satisfied in all coordinate systems in this case. An example of anisotropy with a triple degeneracy is the medium with elastic constants a34 = a35 = a45 = 0 and a33 = a44 = a55. The triple acoustic axis is along the x3 axis. The axis is multiple and thus unstable.
The solutions of eqs (2a)–(c) satisfying the conditions of eqs (3), (4) and (7) should not be eliminated as proposed incorrectly by Vavryčuk (2005). Hence, the correct equations for determining the spurious directions are as follows:
(8a)
or
(8b)
or
(8c)
or
(8d)

However, it should also be pointed out that the directions calculated from eqs (2a)–(c) can coincide with acoustic axes only under additional strict conditions imposed on the elastic constants. These conditions are generally not obeyed in triclinic media (see also Fedorov 1968). Moreover, although the triple degeneracy is physically possible, it is rather rare and not expected to exist in natural materials. At least, no report on the discovery of such materials could be found so far. As confirmed by Vavryčuk, all numerical experiments and conclusions published in Vavryčuk (2005) remain valid despite the reported bugs.

References

Fedorov
F.I.
,
1968
.
Theory of Elastic Waves in Crystals
,
Plenum Press
,
New York
, p.
105
.

Khatkevich
A.G.
,
1963
.
Acoustic axes in crystals
,
Sov. Phys. Crystallogr.
,
7
,
601
604
.

Vavryčuk
V.
,
2005
.
Acoustic axes in weak triclinic anisotropy
,
Geophys. J. Int.
,
163
,
629
638
.